初始化项目，由ModelHub XC社区提供模型

Model: HuggingFaceH4/starchat-beta Source: Original Platform
2026-05-11 15:25:20 +08:00
commit 477daa54ac
33 changed files with 148717 additions and 0 deletions
--- a/trainer_state.json
+++ b/trainer_state.json
@@ -0,0 +1,145 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 5.901639344262295,
+  "global_step": 90,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.07,
+      "learning_rate": 0.0,
+      "loss": 1.8309,
+      "step": 1
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 2e-05,
+      "loss": 1.5321,
+      "step": 8
+    },
+    {
+      "epoch": 0.98,
+      "eval_loss": 1.2855817079544067,
+      "eval_runtime": 6.6367,
+      "eval_samples_per_second": 30.437,
+      "eval_steps_per_second": 1.055,
+      "step": 15
+    },
+    {
+      "epoch": 1.05,
+      "learning_rate": 2e-05,
+      "loss": 1.35,
+      "step": 16
+    },
+    {
+      "epoch": 1.57,
+      "learning_rate": 2e-05,
+      "loss": 1.2071,
+      "step": 24
+    },
+    {
+      "epoch": 1.97,
+      "eval_loss": 1.2619894742965698,
+      "eval_runtime": 5.1554,
+      "eval_samples_per_second": 39.183,
+      "eval_steps_per_second": 1.358,
+      "step": 30
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 2e-05,
+      "loss": 1.1502,
+      "step": 32
+    },
+    {
+      "epoch": 2.62,
+      "learning_rate": 2e-05,
+      "loss": 1.0162,
+      "step": 40
+    },
+    {
+      "epoch": 2.95,
+      "eval_loss": 1.285272240638733,
+      "eval_runtime": 5.1992,
+      "eval_samples_per_second": 38.852,
+      "eval_steps_per_second": 1.346,
+      "step": 45
+    },
+    {
+      "epoch": 3.15,
+      "learning_rate": 2e-05,
+      "loss": 0.9511,
+      "step": 48
+    },
+    {
+      "epoch": 3.67,
+      "learning_rate": 2e-05,
+      "loss": 0.8484,
+      "step": 56
+    },
+    {
+      "epoch": 4.0,
+      "eval_loss": 1.3274288177490234,
+      "eval_runtime": 5.1899,
+      "eval_samples_per_second": 38.922,
+      "eval_steps_per_second": 1.349,
+      "step": 61
+    },
+    {
+      "epoch": 4.2,
+      "learning_rate": 2e-05,
+      "loss": 0.7971,
+      "step": 64
+    },
+    {
+      "epoch": 4.72,
+      "learning_rate": 2e-05,
+      "loss": 0.6981,
+      "step": 72
+    },
+    {
+      "epoch": 4.98,
+      "eval_loss": 1.3993656635284424,
+      "eval_runtime": 5.213,
+      "eval_samples_per_second": 38.749,
+      "eval_steps_per_second": 1.343,
+      "step": 76
+    },
+    {
+      "epoch": 5.25,
+      "learning_rate": 2e-05,
+      "loss": 0.6462,
+      "step": 80
+    },
+    {
+      "epoch": 5.77,
+      "learning_rate": 2e-05,
+      "loss": 0.5668,
+      "step": 88
+    },
+    {
+      "epoch": 5.9,
+      "eval_loss": 1.4719988107681274,
+      "eval_runtime": 5.1996,
+      "eval_samples_per_second": 38.849,
+      "eval_steps_per_second": 1.346,
+      "step": 90
+    },
+    {
+      "epoch": 5.9,
+      "step": 90,
+      "total_flos": 383994839433216.0,
+      "train_loss": 0.9728257921006944,
+      "train_runtime": 2307.7787,
+      "train_samples_per_second": 10.108,
+      "train_steps_per_second": 0.039
+    }
+  ],
+  "max_steps": 90,
+  "num_train_epochs": 6,
+  "total_flos": 383994839433216.0,
+  "trial_name": null,
+  "trial_params": null
+}